The mid-presentation will take place February 25 at 9:15 at INRIA. The real ^mis-presentation will take place in 15 days at Paris.
The title of the mid-thesis presentation is « Variable speed limits in a large-scale urban network for improved environmental sustainability ».
Abstract :
" The problem of urban traffic management for improved environmental sustainability is addressed. The objective of this kind of traffic management is to reduce the energy consumption (electricity or fuel), as well as the emission of pollutants, without deteriorating the traffic performance of the urban road network under consideration. Eco-management can correspond either to the control of vehicles or to the control of infrastructures. The vehicle-side control includes eco-driving and eco-routing control designs. Eco-driving consists in computing a vehicle trajectory that minimizes the emissions or energy consumption along a given route, under technical and environment constraints. Eco-routing consists in planning an emission or energy-minimal route, given an origin and a destination. The more connected the vehicles, the more efficient the eco-driving and eco-routing algorithms as their prediction is more accurate. The infrastructure-side control corresponds to the dynamic management of traffic actuators such as speed limits and traffic light signals. Its objective is to reduce the total pollutant emissions and energy consumption in the network. On a macroscopic level, this approach is usually much faster to compute because it is based on coarser traffic models. Moreover, some frameworks such as decentralized control can contribute to the reduction of computation times in large-scale networks. Hence, we choose to focus on the infrastructure-side control in this work. The contributions of this work to the traffic eco-management, through the control of actuators at the infrastructure level, may be summarized as follows. Firstly, a complete formalization of a macroscopic traffic model adapted to the urban environment is proposed. It includes a methodology for treating intersections with -- and without -- traffic light signals, as well as a procedure to model multi-lane roads. This traffic model is associated with a revisited macroscopic emission and energy consumption model. Then, a control design based on variable speed limits for improved environmental sustainability is proposed. This actuator directly impacts the energy consumption and pollutant emissions as speed limits affect the accelerations and average speeds through the network. Lastly, experiments are conducted to control a synthetic urban road network with the proposed controller. They are performed by considering a system characterized by a macroscopic traffic model in a first set of simulations -- and by a microscopic traffic model, which is more precise and close to reality, in a second set of simulations. The results are analyzed so as to evaluate the impact of the controller. »